The impact of acute alcoholic hepatitis in the explanted recipient liver on outcome after liver transplantation

Authors

  • Jennifer T. Wells,

    1. Section of Gastroenterology and Hepatology, Department of Medicine, University of Wisconsin School of Medicine and Public Health, Madison, WI
    Search for more papers by this author
  • Adnan Said,

    1. Section of Gastroenterology and Hepatology, Department of Medicine, University of Wisconsin School of Medicine and Public Health, Madison, WI
    Search for more papers by this author
  • Rashmi Agni,

    1. Department of Pathology, University of Wisconsin School of Medicine and Public Health, Madison, WI
    Search for more papers by this author
  • Santiago Tome,

    1. Liver Unit Internal Medicine Department. Hospital Universitario de Santiago de Compostela. Spain
    Search for more papers by this author
  • Sarah Hughes,

    1. Department of Pathology, University of Wisconsin School of Medicine and Public Health, Madison, WI
    Search for more papers by this author
  • Parul Dureja,

    1. Section of Gastroenterology and Hepatology, Department of Medicine, University of Wisconsin School of Medicine and Public Health, Madison, WI
    Search for more papers by this author
  • Michael R. Lucey

    Corresponding author
    1. Section of Gastroenterology and Hepatology, Department of Medicine, University of Wisconsin School of Medicine and Public Health, Madison, WI
    • Section of Gastroenterology and Hepatology H6/516 CSC, University of Wisconsin Hospital and Clinics, 600 Highland Avenue, Madison, WI 53711
    Search for more papers by this author
    • TEL: (608) 263-7322; FAX: (608) 265-5677


Abstract

Patients with clinical acute alcoholic hepatitis (AAH) are not considered suitable candidates for orthotopic liver transplantation (OLT). The histological correlates of AAH are often seen in the explanted liver at the time of transplantation. The importance of these findings remains inconclusive regarding their role as a prognostic marker for patient or allograft health. Our aim was to examine the explanted liver of patients with purely alcoholic liver disease (ALD) for findings of histologic AAH and to correlate these to patient and graft outcomes. We compared patients with and without histological AAH with patients transplanted for non-ALD. Of 1,097 liver transplant recipients, 148 had ALD and 125 were non-ALD control patients with similar demographics. Thirty-two of 148 ALD patients had histologic AAH, and 116 had bland alcoholic cirrhosis (BAC). Twenty-eight percent of the ALD patients reported <6 months abstinence, and 54% reported <12 months abstinence. There was a statistically significant relationship between the presence of histologic AAH and abstinence durations <12 months (P = 0.009), but not <6 months. Overall, posttransplantation patient and graft survival between the ALD and non-ALD groups was not significantly different (P = 0.53). Furthermore, patient and graft survival between ALD patients with histologic AAH and BAC were similar (P = 0.13 and P = 0.11, respectively). The rate of posttransplantation relapse among ALD patients was 16%; however, there was no increase in graft loss, nor was there decreased survival compared with controls. The patients with histologic AAH and those with BAC had no differences in posttransplantation relapse (P = 0.13). In multivariate analysis, patient and graft survival was not influenced by pretransplantation abstinence or posttransplantation relapse. In conclusion, histological alcoholic hepatitis in the explant did not predict worse outcome regarding relapse, and allograft or patient survival for liver transplant recipients. Caution should be exercised when liver histology is used to discriminate among suitable candidates for OLT concerning alcoholic patients. Liver Transpl 13: 1728–1735, 2007. © 2007 AASLD.

Acute alcoholic hepatitis (AAH) refers to both a clinical syndrome and a set of histological liver biopsy appearances. The syndrome of AAH always arises after sustained consumption of alcohol. Its clinical manifestations cover a wide spectrum, ranging from asymptomatic jaundice to a fatal combination of fever, jaundice, coagulopathy, and leukocytosis.1 The histological counterpart of AAH is characterized by polymorphonuclear infiltrates, centrilobular hepatocyte swelling and degeneration, macrovesicular and microvesicular steatosis, Mallory bodies, and pericentral-perisinusoidal fibrosis.2

Typically, patients with life-threatening clinical AAH are managed medically; they are not considered suitable for orthotopic liver transplantation (OLT). The eligibility for OLT of ALD patients has been questioned both within and outside the transplant community as a result of the stigma associated with alcohol abuse, the notion that alcoholic patients are responsible for their illness, and the observation that many alcoholic patients resume drinking alcohol at some point after transplantation.3 As part of this ongoing dialogue, there has been a widespread consensus that patients with life-threatening AAH should be required to wait through a 6-month period of abstinence from alcohol.4 One justification for a required period of abstinence has been that it gives the native liver a chance to recover, and thereby avoid unnecessary transplantation.5 However, because severe AAH has a high mortality, even with maximal medical assistance, we have asked whether there is a population for patients with AAH in whom transplantation is an appropriate treatment.6

To date, although there have been occasional reports of liver transplantation in the setting of clinical AAH, there have been no systematic studies of liver transplantation in patients with AAH unresponsive to medical management.7 An alternative methodology to address the effect of AAH on outcome after OLT would be to use the histological features of AAH as a surrogate for the clinical syndrome. By means of this approach, Tome et al.8 found in a Spanish cohort that histological AAH did not influence posttransplantation drinking behavior or clinical outcome. In contrast, Conjeevaram et al.9 reported that 4 of 8 ALD patients with steatosis and Mallory bodies in their explanted livers returned to heavy drinking, compared with only 2 (4%) of 49 without these histologic findings. Here, we consider the effect, if any, of histological AAH on transplantation outcome in a large U.S. cohort of patients transplanted for ALD uncomplicated by concomitant hepatitis infection.

Abbreviations

AAH, acute alcoholic hepatitis; OLT, orthotopic liver transplantation; ALD, alcoholic liver disease; AODA, alcohol and other drugs of abuse; MELD, Model for End-Stage Liver Disease; BAC, bland alcoholic cirrhosis.

METHODS

Patients

We used a prospectively maintained liver transplant database to identify all subjects transplanted at the University of Wisconsin School of Medicine and Public Health between January 1984 and December 2002 for a “sole diagnosis” of alcoholic liver disease (ALD). One thousand ninety-seven patients 18 years old or older underwent OLT during the 18-year study period (Fig. 1). From this cohort, 210 patients were found to have a diagnosis of ALD. A detailed chart review was performed to confirm the diagnosis, and we thereby limited our final study group to 148 ALD patients (106 men, 42 women). The diagnosis of ALD was made in patients with clinical, biochemical, radiographic, and often pathologic evidence of cirrhosis in addition to a history of marked, regular, and heavy alcohol consumption for at least 5 years (men, >60 g per day; women, >40 g per day). The remaining 62 patients were excluded as a result of additional diagnoses, including coexistent causes of liver disease. We constructed a control group comprising 125 non-ALD liver transplant recipients (76 men, 49 women) matched for year of transplantation. These patients had liver disease attributed to: alpha-1-antitrypsin deficiency (n = 3), fulminant liver failure (n = 3), autoimmune hepatitis (n = 5), Caroli disease (n = 1), cholangiocarcinoma (n = 1), cryptogenic (n = 18), hepatitis C (n = 20), hemochromatosis (n = 3), primary biliary cirrhosis (n = 17), polycystic liver disease (n = 2), primary sclerosing cholangitis (n = 24), sarcoidosis (n = 1), secondary biliary cirrhosis (n = 1), nonalcoholic steatohepatitis (n = 6), Wilson disease (n = 3), hepatitis B (n = 6), and miscellaneous (n = 11).

Figure 1.

Flow chart of methods for patient selection.

Patients with known hepatocellular carcinoma and those in whom hepatocellular carcinoma was incidentally discovered during dissection of the explanted liver were included in this study and categorized by the primary cause of liver disease. Liver transplantation was performed in an elective (nonurgent) protocol in all cases, with the exception of patients with fulminant liver failure. Standard procedure in all cases included OLT with piggyback technique, and primary immunosuppression with calcineurin inhibitors (cyclosporine or tacrolimus) and corticosteroids. During the period of study, patients with clinically overt signs and symptoms of AAH (recent alcohol use defined as within 6 weeks of clinical presentation, jaundice, fever, and leukocytosis) were not considered for liver transplantation.

Clinical and Alcohol Evaluation

Drinking behavior was assessed by reviewing the clinical records, including pretransplantation evaluations by our AODA (alcohol and other drugs of abuse) counselors. In particular, we documented previous treatment of alcoholism, and evidence for and character of alcoholic relapse after transplantation. We identified variables that included amount of alcohol intake per day, duration of alcohol use, and pretransplantation evaluation and treatment of alcoholism. A 6-month pretransplantation abstinence period has been the standard before placing alcoholic patients on the transplant list at our institution; however, a more lenient standard was sometimes accepted, particularly in the early years of the transplant program. The pretransplantation duration of abstinence used in the present study was based on documented history from patients and/or family members. Alcoholic relapse in this study was defined as any posttransplantation recorded use of alcohol documented from the patient or family members, or by discovery of positive blood alcohol levels when the patient presented for outpatient or inpatient care.

Assessment of Liver Function and Clinical Severity of Liver Disease

A calculated Model for End-Stage Liver Disease (MELD) score was used to assess global liver function in all patients before liver transplantation. In addition, among those patients with alcohol-induced liver cirrhosis, the Maddrey discriminate function was calculated. All scores were obtained from laboratory studies undertaken on samples taken immediately before transplantation.

Liver Pathology

Two independent pathologists (R.A., S.H.) blinded to the patients' diagnoses analyzed the slides of the explanted livers in both the ALD and control groups. Liver cirrhosis was confirmed in all cases. A diagnosis of superimposed alcoholic hepatitis was made when at least 3 of the following criteria were prominently present: ballooning degeneration, infiltrates of neutrophils, Mallory hyaline degeneration, and liver steatosis.2

Follow-up and Survival

All patients were followed regularly in our transplant clinic. Data of mortality and survival were available in all with the exception of 6 ALD and 7 control cases. The posttransplantation follow-up period was tracked through December 2003 with a mean among survivors of 105.0 months (range, 5-205 months) in the whole series. The mean follow-up in the group of patients with ALD alone was 90.5 months compared with that of the non-ALD control group, which was 74.6 months.

Ethical Issues

This study was approved by the Institutional Review Board of the University of Wisconsin Hospitals and Clinics.

Statistical Analysis

Continuous variables were compared by t tests or Mann-Whitney tests with variances reported as standard deviations. Categorical variables were compared with χ2 or the Fisher exact test. Patient and graft survival were analyzed by the Kaplan-Meier method, and survival curves were compared by means of the log rank nonparametric test. A 2-sided P value of <0.05 was considered statistically significant. Predictors of survival that were significant in univariate analyses were entered into a multivariate Cox proportional hazard model to assess independent predictors of survival.

RESULTS

Patient Population

Of 1,097 liver transplant recipients, 148 had ALD and 125 were non-ALD control patients (Table 1). Most patients were white men who were an average age of 50 years. The average MELD score was 19.8. There was no statistically significant difference in the demographics between the ALD and control groups. Pretransplantation AODA counseling was integrated into the transplant evaluation in 53% of ALD patients. Further analysis showed that AODA evaluation was performed in 26 of 81 patients in the early years of the program, whereas since 1993, 52 of 57 received AODA evaluation (P < 0.0001). None of the control patients demonstrated regular alcohol use, abuse, or dependence issues at pretransplantation evaluation.

Table 1. Patient Characteristics
CharacteristicALDControl (non-ALD)P value
  1. Abbreviation: ALD, alcoholic liver disease; MELD, Model for End-Stage Liver Disease; HBV, hepatitis B virus; HCV, hepatitis C virus; NASH, nonalcoholic steatohepatitis.

N148125 
Age (yr), mean (SD)50.27 (9.34)52.16 (8.99)0.10
Gender (M:F)106:4276:490.07
Race139 white, 9 other113 white, 12 other0.36
MELD score, mean (SD)19.84 (9.66)18.58 (10.87)0.31
Cold ischemia time (h), mean (SD)11.53 (8.48)13.53 (16.29)0.22
Cause of liver disease (n)   
 HBV8 
 HCV28 
 Cholestasis37 
 NASH6 
 Other metabolic6 
 Fulminant5 
 Other35 

Analysis of Pathology

Of the patients undergoing liver transplantation with the diagnosis of ALD alone, the explanted liver of 32 patients (22%) fulfilled the diagnostic criteria for histologic AAH. The remaining 116 ALD patients had bland alcoholic cirrhosis (BAC). The demographics of these ALD subgroups were similar in regards to mean age, race, sex distribution, and cold ischemia time (Table 2). Similarly, there was no difference between the pretransplantation severity of disease among the alcoholics with histologic AAH, those with BAC, and the control group on the basis of the calculated MELD scores. Of the 148 ALD patients, the mean Maddrey degree of freedom was 37.83 in the histologic AAH compared with a mean of 32.48 in the BAC group (P = 0.32).

Table 2. Demographics and Results of Patients With Histologic Acute Alcoholic Hepatitis (AAH) and Cirrhosis vs. Bland Alcoholic Cirrhosis (BAC)
CharacteristicAAH (n = 32)BAC (n = 116)P value
  • Abbreviations: AODA, alcohol and other drugs of abuse; ETOH, ethyl alcohol consumption; MELD, Model for End-Stage Liver Disease.

  • *

    Early relapse indicates alcohol use within 1 year after transplantation; marked relapse, ≥6 alcoholic drinks in one day.

Age (yr)50.1150.380.87
Gender (M:F)22:1083:330.83
Race31 white, 1 other108 white, 7 other1.00
Body mass index (kg/m2), mean (SD)25.61 (5.90)26.63 (2.97)0.38
Cold ischemia time (h), mean (SD)11.99 (9.44)9.75 (2.99)0.19
AODA (γ/N)26/3252/1160.001
ETOH treatment (n)32190.001
MELD score, mean (SD)20.56 (9.74)19.68 (9.71)0.65
Maddrey discriminate function (mean)37.8332.480.32
Abstinence before transplantation  0.0001
 Average (mo) (range)9.72 (0-18)25.98 (0-180) 
 0-6 months (n)734 
 6-12 months (n)1721 
 >12 months (n)861 
Relapse*10140.13
 Early (n)221.00
 Marked (n)8120.26
 Time to relapse (mo) (range)45.44 (12-99)38.93 (3-132)0.64

Pretransplantation Abstinence and Degree of Histologic Liver Disease

The mean abstinence period in the ALD group was 21.89 months (range, 0-180 months). Twenty-eight percent (41 of 148) of these patients were reported to have been abstinent for <6 months before transplantation. Thirty-eight (26%) and 69 (46%) ALD patients had abstinence intervals of 6-12 months and >12 months, respectively (Fig. 2). There was a statistically significant relationship between histologic AAH and recorded abstinence, with significantly more AAH patients than BAC patients having abstinence durations of <12 months (P = 0.009) (Table 3). Thirty-four percent of the ALD patients had completed some form of alcohol treatment or counseling before transplantation. All of the patients subsequently found to have histological AAH had reported alcoholism treatment before transplantation. Thirty of 32 patients with histological alcoholic hepatitis were abstinent for at least 4 weeks before transplantation.

Figure 2.

Pretransplantation duration of abstinence. Breakdown of patients with ALD according to their pretransplantation abstinence interval.

Table 3. Histologic Changes Associated With Pretransplantation Abstinence Intervals of 6 and 12 Months
IntervalAAH (n)BAC (n)P value*
  1. Abbreviations: AAH, acute alcoholic hepatitis + cirrhosis; BAC, bland alcoholic cirrhosis).* Odds ratio, 3.21; 95% confidence interval, 1.33-7.74.

6 months  0.51
 <6 months735 
 >6 months2581 
12 mo  0.009
 <12 months2456 
 >12 months860 

Analysis of Survival

The overall posttransplantation patient and graft survival between the ALD cohort and non-ALD controls was not significantly different (Figs. 3 and 4). The ALD group had a median survival of 112 months with 81 deaths and 87 graft losses. When compared with the control group, the median survival of which was 111 months with 52 deaths and 52 graft losses, there was no significant difference (P = 0.53). The subset of ALD patients with histological AAH and BAC were compared for patient and graft survival, and there were no differences (P = 0.13 and P = 0.11 respectively) (Fig. 5 and 6). In multivariate analysis, neither the duration of abstinence nor the occurrence of relapse influenced patient (Table 4) or graft survival. The causes of mortality were similar between the ALD and control groups (Table 5).

Figure 3.

Posttransplantation survival of the ALD patients vs. non-ALD (control) patients.

Figure 4.

Graft survival of ALD patients vs. non-ALD (control) patients. The numbers of patients at each time point are shown.

Figure 5.

Patient survival for patients with histologic AAH vs. BAC. The numbers of patients at each time point are shown.

Figure 6.

GRAFT survival for patients with histologic AAH vs. BAC. The numbers of patients at each time point are shown.

Table 4. Multivariable Analysis of Patient Survival
VariableOdds ratio (95% CI)P value
  1. Abbreviations: 95% CI, 95% confidence interval; MELD, Model for End-Stage Liver Disease; AAH, acute alcoholic hepatitis + cirrhosis; BAC, bland alcoholic cirrhosis.

Age (yr)1.02 (0.99-1.06)0.14
MELD score0.97 (0.94-1.00)0.05
AAH vs. BAC0.59 (0.31-1.10)0.10
Posttransplantation alcoholic relapse1.56 (0.85-2.89)0.16
Duration of abstinence before transplantation0.99 (0.98-1.01)0.33
Table 5. Causes of Death After Transplantation
Cause of deathAlcoholic liver disease (n)Control (n)
Cardiovascular1411
Malignancy1514
Infection1811
Recurrent cirrhosis/liver failure32
Other248
Total7446

Posttransplantation Alcohol Use

The rate of documented posttransplantation alcoholic relapse among the ALD patients was 24 (16%) of 148. Among ALD patients who relapsed to alcohol use, there was no increased incidence of graft loss, nor was there decreased survival when compared with the non-ALD group. The posttransplantation relapse rate of ALD patients with histological AAH was 10 (31%) of 32, for which 8 of 10 subjects reported harmful drinking. The patients with BAC had posttransplantation recurrence of alcohol use documented in only 14 (12%) of 118. This difference did not reach statistical significance (P = 0.13).

DISCUSSION

ALD is the leading cause of cirrhosis and the second most common reason for liver transplantation in the United States.10 Despite the continued controversy regarding the candidacy of patients with alcohol-induced liver disease, most, although not all, studies report that early- and intermediate-term outcome of liver transplantation in patients with ALD is as good as that of patients undergoing transplantation for other causes of liver disease.11–14 The most recent guidelines of the American Association for the Study of Liver Disease state that rejection, graft failure, and the need for retransplantation all are less common in patients with ALD compared with patients transplanted for other conditions, with the exception of the increased incidence of pharyngeal, esophageal, and gastric malignancies in patients with ALD.5 Concern for posttransplantation alcoholic relapse leading to decreased patient and/or graft survival remains at the heart of this debate.15 Pfitzmann et al.16 recently published data suggesting that relapse to addictive drinking after OLT is associated with worse long-term survival.

The syndrome of AAH spans a wide clinical and histological spectrum, and when superimposed on cirrhosis, it is well recognized to predict a guarded prognosis, as shown by the VA Cooperative Study Group in 1991.17 Unfortunately, even when patients adhere to all aspects of medical management, including strict abstinence, recovery is not guaranteed, and many patients with severe AAH die, with the onset of renal failure being a particularly ominous indicator.18, 19 The role of liver transplantation in this subset of patients is a polarizing debate. The main concern is that patients with AAH who have invariably recently consumed alcohol are at a higher risk for alcoholic relapse after liver transplantation that could impair patient and graft survival.3

In addition to the standard assessment of medical and surgical suitability common to all candidates for liver transplantation, the evaluation of patients with ALD includes identification of the risk that the patient will relapse to heavy and deleterious alcohol use after the transplantation.20 An experienced AODA counselor is a vital member of the transplant teams both before and after transplantation, and AODA assessments that voice concerns should be strongly considered.21 The use of AODA in the first 10 years of our transplant program was initially quite low at 32%. In the more recent era, 1993-2003, the importance of the AODA expertise in a multidisciplinary evaluation has been recognized, and approximately 92% of subjects transplanted for ALD alone were assessed by AODA.

The term “acute alcoholic hepatitis” is frequently used interchangeably in reference to the previously described clinical and histological entities, when in fact they often have very different clinical courses and consequences. Recovery from clinical AAH is dictated by abstinence from alcohol and the severity of the clinical syndrome and is often greatly improved within several weeks discontinuing alcohol. The duration of the histological changes of AAH in abstinent patients, however, remains uncertain. One small study showed persistent histologic changes of alcoholic hepatitis in 6 of 11 patients for up to 3 years after cessation of drinking.22 The triad of lobular leukocytosis, steatosis, and Mallory hyaline in patients claiming to be abstinent has been considered as a surrogate marker of clandestine drinking.23 These histological features may be seen in patients with nonalcoholic steatohepatitis. There was, however, no difference in the body mass indices in patients with histological alcoholic hepatitis and those with BAC, making it unlikely that body mass index differences alone account for the histological changes seen.

In the present study, we used retrospective reexamination of explanted liver histology to investigate the incidence of histologic AAH and its relationship with drinking patterns before and after transplantation, as well as posttransplantation graft and patient survival.

The proportion of our alcoholic cohort in whom histological AAH was identified in the explants (22%) was less than in other published studies.7, 8 For example, Tome et al.8 reported that 54% of patients in a Spanish cohort of alcoholics undergoing liver transplantation had histological AAH in the explanted liver. These data might represent a more stringent discrimination against actively drinking patients in our program than in the Spanish program. In our study, the presence of histological acute hepatitis correlated with a pretransplantation abstinence period of <12 months, but not 6 months. Given these findings, we cannot say whether histological AAH represents clandestine drinking, or whether they are merely residual histological markers. In a subsequent prospective study, we observed some alcohol use in 25% of alcoholic subjects awaiting liver transplantation (M.R.L., unpublished observations). Consequently, it is likely that at least some of our subjects with histological AAH had been drinking recently. Importantly, our patients with histological AAH had similar posttransplantation survival rates when compared with ALD patients with BAC and with the non-ALD control groups. These data corroborate a report from Tome at el.8 from Spain but are in contrast to the smaller study of Conjeevaram et al.,9 wherein 4 of 8 patients with steatosis and Mallory bodies in their native livers returned to heavy drinking compared with only 2 (4%) of 49 without these histologic findings (P < 0.05).

In our study, 16% of the ALD patients acknowledged alcohol use during routine posttransplantation follow-up. This is at the lower end of the range of other published rates, which have reported relapse rates of 8%-95%.9, 24–26 Within our relapse data analysis, surprisingly few ALD patients admitted to occasional minor drinking followed by restored abstinence, often referred to as a “slip.” Much more common was the admission of heavier alcohol ingestion. This is in contrast to the recent prospective study by DiMartini et al.,25 in which both slips and episodes of harmful drinking were systematically recorded. Whereas alcohol use was found in >40% of subjects, slips were by far the most common event. Thus, it is likely that our retrospective methodology has failed to capture minor drinking. Put another way, the clinical circumstances of the present study favored concealment of minor amounts of drinking, whereas heavier use was harder to hide.

Although many studies have sought to elucidate the most critical risk factors predictive of relapse,16, 21, 27, 28 consistent and reproducible data are elusive. Despite the lack of definitive studies, almost all U.S. transplant programs and insurance companies have adopted a 6-month abstinence period as a requirement for ALD patients.4 There is an inherent difficulty in this: patients recognize that it is not in their interest to admit alcohol use, so they conceal it, and few objective data can confirm or refute their claims of abstinence. We have speculated that this resistance to open admission of alcohol use may be a serious hindrance to patient recovery.29 In our study, there was no statistically significant correlation between patients who experienced relapse into alcohol use and the pretransplantation declared abstinence period. Similarly, the presence of histological AAH was not associated with the likelihood of alcoholic relapse after transplantation.

We conclude that patient and graft survival among those transplanted patients with alcoholic cirrhosis is similar to those transplanted for other indications. Additionally, survival was also similar between the subset of patients with superimposed histologic AAH vs. those with BAC. Although the histological manifestations of AAH were correlated with abstinence time periods of <12 months, there was no increase in alcoholic relapse in those subjects with prior histological AAH. Our data imply that results of a pretransplantation biopsy showing histological AAH should not be interpreted, a priori, as an impediment to transplant candidacy. These findings should be confined to those patients with ALD and superimposed histologic hepatitis. The suitability of patients with clinical AAH as candidates for liver transplantation will need to be addressed in further studies.

Ancillary